The present invention generally relates to the field of information handling systems, and particularly to memorialization of conversations.
In the modem business environment, it is often desirable to memorialize conversations and meetings for later referral and for record keeping purposes. Typically, the audio portion of a conversation is recorded onto magnetic tape with a tape recorder, or video and audio portions of a conversation may be recorded in instances where it is desired to memorialize visual aspects of a meeting. Often, a scribe or secretary may be present at the meeting to manually capture the conversation to create a text transcript of the meeting. This is done to further the objective of associating speakers with uttered passages of conversation so that a person reading the transcript at a later time would be able to better understand the conversation by knowing which speaker spoke which utterances. However, using a human being to create a transcript in real time is inefficient and results in errors since conversations typically occur at a greater rate than the rate at which a human being can write down or type the conversation information. Problems increase when the person doing the transcribing is required to participate in the conversation. In addition, an operator is required to operate recording equipment such as microphones and video cameras requiring an additional human being who is not able to fully participate in the conversation while operating the equipment. Thus, there lies a need for a system that is capable of automatically recording and transcribing a meeting and which is further capable of automatically operating meeting recording devices.
The present invention is directed to a method for memorializing a conversation of a plurality of speakers. In one embodiment, the method includes steps for sampling a sample utterance of each of the plurality of speakers thereby producing a plurality of sample utterances each having a characteristic corresponding to a corresponding one of the plurality of speakers, associating a characteristic of the sample utterances with the corresponding one of the plurality of speakers, recording the conversation by saving the conversation to a storage medium as the conversation occurs whereby conversation utterances generated by each of the plurality of speakers during the conversation are saved to the storage medium, identifying the one of the plurality of speakers who generated each one of the conversation utterances by matching the characteristic of the sample utterance associated with the corresponding one of the plurality of speakers with a characteristic of the conversation utterance, and associating information regarding the identified speaker with at least one of the conversation utterances.
The present invention is further directed to an information handling system for memorializing a conversation of a plurality of speakers. In one embodiment, the information handling system includes a processor for executing a program of instructions on the information handling system, a memory coupled to the processor for storing the program of instructions executed by the processor, and a transducer coupled to the information handling system and controlled by the processor for converting utterances of the plurality of speakers into a signal interpretable by the information handling system wherein the information handling system generates a transcript of the conversation, the transcript including information associated with each of the utterances, the information regarding the corresponding speaker of each utterance.